Skip to main content


See other formats



Muhaimad, A Alaraj 
Bachelor of Electrical Engineering, Qassim University, 2008 

A thesis submitted to the 
Faculty of the Graduate School of the 
University of Colorado in partial fulfillment 
of the requirements for the degree of 
Master of Science 
Electrical Engineering 

This thesis for the Master of Science degree by 
Muhannad, A Alaraj 
has been approved for the 
Electrical Engineering Program 

Jae-Do Park, Chair 
Tim C. Lei 
Jason Ren 

Muhannad, A Alaraj (M.S., Electrical Engineering) 

Energy Harvesting From Microbial Fuel Cell Using Self-Synchronous Flyback Converter 
Thesis directed by Assistant Professor Jae-Do Park. 


Microbial Fuel Cells (MFCs) use biodegradable matter, such as wastewater and 
animal droppings to generate electrical energy. To harvest the energy from MFC, power 
electronic converters have recently been used because of their advantages, such as the 
ability to store the harvested energy and the ability to control MFC voltage. Although 
power electronic converters have advantages to be used to harvest the energy, the diode 
based energy harvesters suffer from the low efficiency because of the diode losses. 
Replacing the diode with a MOSFET reduces the loss because MOSFET have lower 
conduction loss, but this replacement causes the synchronous MOSFET to be floating, 
which requires an isolated gate signal. This study presents harvesting energy from MFC 
using self-synchronous flyback converter, which improved the harvesting efficiency by 
37.6 % compared to a diode based boost converter. 

The form and content of this absfract are approved. I recommend its publication. 

Approved: Jae-Do Park 



I dedicate this work to Abrar, my lovely wife, who has always believed 
and supported me. 



This thesis would not have been possible without the generous support of Qassim 





Introduction 1 

Electrical Characteristics of MFC 1 

Voltage-Current Polarization Curve 1 

Internal Resistance 3 

Maximum Power Point 3 

Advantages and Disadvantages 4 


Intorduction 5 

Passive Energy Extraction 5 

Resistors 5 

Supercapacitors 6 

Charge Pumps 7 

Active Energy Extraction 8 

Power Electronics Converters 8 

Boost Converter 9 


Introduction 12 

Basic Topology for Flyback Converter 12 

Operation of the Flyback Converter 13 

Continuous versus Discontinuous Flux Operation 14 

S5aichronous Flyback Converter 16 



Introduction 17 

Self-Sjaichronous Flyback Converter 18 

Operation of the Self- Synchronous Flyback Converter 18 

Hysteresis Controller 22 

System Simulation 23 

Overall System 24 


Self-Synchronous Flyback Converter 26 

System Parameters 26 

Filtering the MFC Voltage Ringing 30 

Results 32 

Calculations 34 

Boost Converter 38 


Comparison 42 

Conclusion 46 






V. 1 Table of Parameters 29 

V.2 Self-S)aichronous FLyback Converter Experiment Results 35 

V.3 Boost Converter Experiment Results 38 




1. 1 Polarization Curve of Two Different MFCs [5] 

1.2 MFC Electrical Equivalent Circuit [5] 

II. 1 Simple Charge Pump 

II. 2 Schematic Diagram of Boost Converter 

III. 1 Flyback Converter 

III. 2 First Mode of Operation 

111. 3 Second Mode of Operation 

111.4 S5aichronous Flyback Converter 

rV. 1 Self-S)Tichronous Flyback Converter 

IV.2 Synchronous Driving Circuit Operation 

IV.3 Non-inverting Hysteresis Controller 

IV. 4 Simulation Results 

IV. 5 Overall System 

V. l Winding Machine 

V.2 Ringing Waveforms 

V.3 Non-inverting hysteresis controller with capacitor at the input 

V.4 Waveforms Afer Filtering 

V.5 Experiment Set 

V.6 Self-Synchronous FLyback Converter Experiment Waveforms. .. 
V.7 Efficiency of the Self- Synchronous Flyback Converter vs. Time. 

V.8 Boost Converter Experiment Waveforms 

V.9 Efficiency of the Boost Converter vs. Time 

VI. 1 MFC Voltage for Both Experiments vs. Time 43 

VI.2 MFC Current For Both Experiments vs. Time 43 

VI.3 Switching Frequency For Both Experiments vs. Time 44 

VI.4 Output Capacitor Voltage for Both Experiments vs. Time 44 

VI.5 Efficiency of Both Converters vs. Time 45 




The main purpose of this thesis is to harvest the energy efficiently from Microbial 
Fuel Cell (MFC). MFC uses biodegradable matter, such as wastewater and animal 
droppings to generate electrical energy. Employing the bacteria to generate electricity is 
the basic idea of the microbial fuel cell. The bacteria oxidize their food source, and 
electrons are produced. When closing the circuit, electrons will circulate and electricity 
will be generated. In the recent researches, the MFC was improved to produce big enough 
power to be cbnsidered as a power source [21, 22]. MFCs can be built in the lab [5, 6] 
also in the ocean [1, 2], which is very useful to have a renewable source of electrical 
power under the ocean water. The U.S. Navy supported some researches on the MFC to 
be able to use it under water to power sensors [1, 2], and it might be used for other 
applications. One important problem of MFC is that it has a low output voltage and it 
cannot be connected in series with other MFCs to have a higher output voltage [3, 4], 
because of their nonlinear behavior. 

Electrical Characteristics of MFC 

Voltage-Current Polarization Curve 

The polarization curve is the raltion between the voltage and the current output of 
the MFC. Different MFCs might have different output power, but usually they have 
similar polirization curves. Figure 1. 1 shows two different MFCs, and it is clear that they 
have the same shape but differemt current and voltage values. 




Figure 1.2 MFC Electrical Equivalent Circuit [5]. 
Internal Resistance 

The voltage of the MFC when its terminals are open is generally 0.7 V. This 
voltage on the terminal of the MFC decreases when an external resistance is connected to 
the MFC, in other words, when a current flows through the MFC. The internal resistance 
of the MFC causes this voltage drop. The equivalent circuit of the MFC is shown in 
Figure 1.2 [5]. The value of the internal resistance varies depending on factors such as 
reactor size and environmental conditions. 

Maximum Power Point 

There is a point of operation where the maximum power can be extracted from 
MFC. The maximum power extraction from MFC happens when an external resistance 
equal to the internal resistance is connected to MFC. The maximum power point can be 
seen in the polarization curves that are shown in Figure I.l. The internal resistance of the 

MFC varies depending on the environmental conditions, but in most cases it can be 
assumed to be constant. Using algorithms that have been developed to track the 
maximum power point, the maximum power can be extracted using a variable resistor or 
using DC-DC converters [6, 7]. 

Advantages and Disadvantages 

The MFC has many advantages that make it an important energy source and 
those advantages can be listed as follows: 

• Sustainable power source. 

• Clean power source from the environment point of view. 

• Direct conversion of substrate energy to electricity [8]. 

• Works in ambient and low temperature [8]. 

• Operates in under water environment. 

However, it also has some disadvantages that effect its operation: 

• Low output voltage. 

• Low output current, few milliamps depending on the size of the MFC. 

• Depends on enviroimiental conditions. 

• Cannot be stacked in series to get higher output voltage, because of their 
nonlinear behaviour [3,4]. 




To use the energy generated by the MFC, an electrical circuit needs to be 
connected to harvest this energy. Different harvesters have been used in the recent years' 
MFC research, including resistors and power electronic converters. This chapter will 
briefly review the previous work that has been done to extract the energy from the MFC. 

MFC energy harvesters can be divided into passive harvesters and active 
harvesters. Each type of harvesters will be briefly discussed. 

Passive Energy Extraction 


Extracting the energy from MFC using an external resistor is the most basic 
technique and has been widely used [12, 23, 24]. When an external resistor is connected 

to MFC, the current will start flowing through the resistor and the MFC voltage is given 

^MFC — ^ ^ext 

where, V^fc is MFC output voltage which is the voltage across the external resistor, / is 
the current passing through that resistor, and Rgxt is the resistance of the external resistor. 
When the current passes through the resistor, the extracted power will be dissipated in the 
resistor. This power dissipation shows the amount of extracted energy: 


^disp — j Pdisp where, Paisp — ^ext 

To dissipate the maximum amount of power on the external resistor, the external 
resistor must be equal to the internal resistance of the MFC. This can be seen in the 
following equations: 

p — j2 n 
^out ~ ' "-ext 

J _ ^mfc 


"•^ out 


_ Vmfc „ 
^OMt ~ /'p 1 p \2 ext 
(."^int "^extJ 

= 0, Rgxt ~ ^int = ^ Rgxt = ^int 

A variable external resistance was tested in [12] with developed perturbation and 
observation (P&O) algorithm to track the maximum power point of the MFC, and the 
algorithm is able to set the external resistance equal to the internal resistance even it is 
changing. The disadvantage of using external resistor is that the extracted energy will 
burned in the resistor, which will not make the energy usable. 


Using a supercapacitor is more useful than using a resistor because the 
supercapacitor stores the energy instead of burning it in the resistor. Supercapacitor is a 
simple way to harvest and store the energy from the MFC, by connecting it in parallel 
with the MFC. In [1, 2, 14], a capacitor and DC-DC converter are used to power wireless 
sensors. Different combinations have been used, but they share the idea of connecting the 








Figure II.l Simple Charge Pump, 

capacitor directly to the MFC. Also, the capacitor was used in [13] to develop a MFC 
tester. To determine the charging and discharging frequency, and the optimum capacity 
of the capacitor for given charging and discharging potentials, and the optimum charging 
potentials when the discharge potentials and capacitor values are given. 

Charge Pumps 

Charge pumps were used to harvest the energy from the MFC. Charge pumps 
basically use capacitors and switches, and the operation of the charge pumps can be 
explained in the simple circuit shown in Figure II. 1 . Two modes of operation are present 
in the charge pumps. The first mode of operation is when the switches and ^3 are 
closed and the switch ^2 is opened. The capacitor is now in parallel with the supply and it 
will start charging to reach the supply voltage following the capacitor voltage-current 


m = c 

After charging the capacitor the switches 5i and ^3 opene and the switch 52 
close, which is the second operation mode. During this mode the capacitor is in series 
with the supply and the output voltage is equal to: 

Vout = Vin + Vc 

Using charge pumps with the MFC has an advantage of being able to harvest the 
energy with higher voltage. In [11], the charge pump was applied directly to the MFC 
and a supercapacitor was connected at the output of the charge pump to store the energy. 
The charge pump is not a preffered choice because of the low efficiency at 16.6% - 
24.4% [11], and the limited controllability. 

Active Energy Extraction 

Power Electronics Converters 

Passive energy extraction from the MFC is not useful, because resistors bum the 
energy, and capacitors can lead the voltage to drop in the MFC. When a supercapacitor is 
connected to MFC, current flows and charges the supercapacitor, and at some point the 
voltage of the supercapacitor will be equal to the voltage of the MFC and the current will 
stop flowing, which will stop harvesting the energy from the MFC. The better way to 
extract the energy from the MFC is by active energy extraction using power electronics 
converters [5, 6, 15, 16, 19]. The energy can be stored in a capacitor and the voltage of 
the MFC can be maintained within the desirable limits when using converters. 
Inductance, duty ratio, and the switching frequency are the elements that affect the 
energy extraction and their effect was investigated in [15]. 


Boost Converter 

The need to increase the voltage of the MFC made the use of the boost converter 
attractive [5, 16, 17, 18, 19]. The boost converter schematic is shown in Figure II.2. On 
the first time period T^, when the switch Qi is closed and the switch Q2 is open, the 
current will flow through the inductor L. The voltage across the inductor will start 
increasing and the current decreases following the basic inductance current voltage 

^ dt 

where, Vi is the voltage across the inductor L, and Ii is the current passing through it. 
During the second time period T2, the switch should be opened and the switch Q2 
should be closed to forward the current to the load. During this time the current will start 
to flow through the switch Q2 to charge the capacitor: 

Vc = + Vl 

where, is the output capacitor voltage and the Vi is the inductor voltage achieved on 
the first time period Switching between these two modes in high frequency will allow the 
energy to be stored in the capacitor with a boosted voltage. The switching fi-equency 
depends on the time spent in the two modes: 


Ps = Tjr , where = Ti + T2 
The output voltage of the boost converter is: 

o ^ ' 

Figure 11,2 Schematic Diagram of Boost Converter, 

where, D is the duty ratio, which is related to the amount of time spent at each period and 
it can be calculated as follows: 


D = — 


where, and T2 are the times spent on the first and second period respectively. Notice 
that the maximum number of the duty ration is one. In [5], a simple boost converter was 
used with a MOSFET as and a diode as Q2. The reported efficiency was low at 43.8%. 
The main reason of this low efficiency is the diode drop, because the voltage across the 
diode is around 0.6 V and the current flows through this diode during the second time 
period. The diode losses can be calculated as foUows: 

Pioss = ^mfc (1 ~ 7)) 


Loss of the diode is very high especially compared to the low power output of 
MFC, which is drawback of the diode-based boost converter. To avoid the high loss of 
the diode, a synchronous boost converter was used in [16]. The synchronous boost 
converter replaces the diode by a MOSFET, because the MOSFET has low on-resistance. 
The problem of using the sjoichronous boost converter is that the MOSFET in place of 
the diode will become a floating switch, which needs to be driven by a separate or 
isolated source. In [16] a transformer was used to drive the synchronous MOSFET by 
isolated signal, and an efficiency of 75.9% has been achieved. 




The idea of using the s5nichronous boost converter with a transformer-based 
circuit to drive the synchronous floating switch makes the flyback converter a viable 
alternative, because the flyback converter already has a transformer that can be used to 
drive the sjoichronous floating switch. Hence, the sjaichronous flyback converter will be 
more efficient than the diode-based boost converter, by eliminating diode and using the 
main transformer for gating signal as well as power transfer. This chapter gives a 
background review on the flyback converter and its operation. 

Basic Topology for Flyback Converter 

Figure III. 1 shows a basic flyback converter schematic. The flyback converter is 
derived fi-om the boost converter, but with a transformer to step up the voltage. The 
transformer is also used to isolate the input and the output, which is required for some 
applications. The transformer must be designed to have a good coupling so that the 
primary and the secondary are linked with minimal leakage flux. The primary and the 
secondary of the transformer windings do not carry current simultaneously. Each side of 
the transformer will carry current only during a part of the switching period depending on 
the duty ratio. 





GATE SIGNAL | -» — I [ 


Figure III.l Flyback Converter, 

Operation of the Flyback Converter 

The operation of flyback converter is defined by two modes: On-State and Off- 
State. Each mode of operation can be described with a separate equivalent circuit that 
will help to understand the operation of the flyback converter. 

When the switch S in Figure III.l closes, the primary winding of the transformer 

is connected to the power supply's positive terminal. During this time the diode on the 

secondary side will be reverse biased and open the secondary side of the transformer. 

Now, the input voltage will appear across the primary winding and the current will flow 

through the primary winding with this current-voltage relation: 

^ dt 

where, Vi is the voltage across the primary inductor and ii is the inductor current passing 
through it. The secondary current will not flow because the secondary circuit is open. 
Hence, the flux will be established in the core by the primary current only. This is the on- 


state mode and Figure III.2 shows the current carrying part of the circuit during this mode 
of operation. The energy will be stored in the magnetic field and it can be calculated 
using this relation: 

_ 1 2 
^stored ~ ^^P^P 

Where Ip denotes the magnitude of the primary current at the end of the conduction 
period. At the end of the first time period the switch S should be opened, which will cut 
the current path on the primary winding. By opening the current path, the voltage of the 
primary winding should be reversed according to the magnetic induction laws. Reversing 
the voltage polarity of the primary side will also reverse the polarity of the secondary 
side. This makes the diode on the secondary side forward biased, which will allow the 
current to pass through the secondary winding and charge the capacitor. This is the 
second mode of operation, and the equivalent circuit can be seen in Figure III.3. 

Continuous versus Discontinuous Flux Operation 

Operating the flyback converter such that the primary switch closes before the 
secondary current goes to zero is known as the continuous flux operation because the 
magentic flux in the transformer core is never zero. In the continuous flux operation, the 
current that flows in the primary winding will not start from zero because of the existing 
magnetic flux. On the other hand, the discontinuous flux operation happen when the 
current at the secondary side of the fransformer goes to zero before the primary switch is 
on. Having zero current in both sides of the transformer means zero flux in the core. 


Ni II N2 


Figure 111,2 First Mode of Operation. 

V - 

V , ^ V =v , 

out jy^ J " sec ^ out 


Figure III.3 Second Mode of Operation. 

Figure III.4 Synchronous Flyback Converter. 

Synchronous Flyback Converter 

The basic flyback converter shares the disadvantage of the diode losses with the 
basic boost converter. For this reason the synchronous flyback converter is better choice 
to increase the efficiency since the MOSFET has much lower conduction losses. The 
synchronous flyback converter, shown in Figure III.4, has same structure as the basic 
flyback converter but with a MOSFET instead of the diode. The challenge with replacing 
the diode with a MOSFET is the driving of the synchronous MOSFET since it is a 
floating switch. The gate drive circuit of the synchronous MOSFET must turn it on when 
the primary switch is off, and must block the reverse current in case of discontinuous 

The reverse current must be considered because the MOSFET conducts 
bidirectional current, and if it is not turned off when the current reduces to zero, the 
energy stored in the capacitor will be discharged through the synchronous MOSFET to 
the transformer. 




The best way to harvest the energy is using power converters, which have many 
advantages over the other harvesters. They are able to maintain the voltage of the MFC at 
certain levels, which will be necessary to extract the maximum power from the MFC, and 
converters give the ability to store the harvested energy. However, their effciency needs 
to be improved as much as possible. 

Since a transformer was used to drive the sjTichronous MOSFET of the boost 
converter because it is a floating switch [6], the idea of using the flyback converter was 
considerable. Therefore the self-sjoichronized flyback converter [20] will be a good 
choice because it synchronizes the secondary MOSFET by itself, so the only thing needs 
to be driven is the primary MOSFET, which is easy to drive. To drive the primary 
MOSFET, a non-inverting hysteresis controller will be used. 

This thesis claims that using the self-synchronized flyback converter will be more 
efficient than using the basic boost converter with a diode because of the MOSFET at the 
secondary side. The operation of the self-synchronized flyback converter and the non- 
inverting hysteresis controller will be discussed in details in this chapter. 


Self-Synchronous Flyback Converter 

The self-s5aichronized flyback converter [20] is basically a sjoichronous flyback 
converter with a designed driving circuit that will drive the sjoichronous MOSFET using 
the voltage across the output capacitor. The self-s5aichronized flyback converter is shown 
in Figure IV. 1, and it can be seen that the sjaichronous MOSFET will be driven by using 
the ou^ut capacitor voltage. Since the capacitor to store the energy must be big enough 
to store the harvested energy, it takes time to build a voltage that can drive the MOSFET. 
At the beginning, when the ou^ut capacitor voltage is zero, the body diode of the 
MOSFET will be used as a switch until the capacitor builds a voltage equal to the 
minimum gate threshold voltage of the MOSFET. This means that at the beginning the 
circuit will act like a basic flyback converter. 

Operation of the Self-Synchronous Flyback Converter 

The operation is similar to the operation of the basic flyback converter discussed 
in the previous chapter. The only thing that will change is when the output capacitor 
builds some voltage. The synchronous MOSFET gate drive circuit will start operating. 
Note that the operation of the flyback converter will not change, so the only thing that 
needs to be discussed is the operation of the synchronous driving circuit. 

The sjoichronous driving circuit [20] consists of two resistors: R[, R'2', and three 
transistors: Qi? Q2, Qs '^ ^ diode D^j. The transistor operates as inverter. The 
transistors Q2 and Q3 operates as push-pull, which is needed to improve the transition of 


Ni N2 9 

Figure IV. 1 Self-Synchronous Flyback Converter, 

the driving circuit. The diode is used to detect the polarity of the voltage across the 
MOSFET, which will be used to operate the transistor Q^. To understand the operation of 
the synchronous driving circuit, its operation will be discussed step by step. 

After opening the primary switch, current starts to flow in the secondary side 
passing through the body diode of the MOSFET. This makes the voltage V^g > 0, which 
will make the diode forward biased with a higher voltage than the base-emitter 
voltage Vijg of the transistor . For this reason the diode wiU be forward biased as in 
Figure IV.2(b). The transistor will remain turned off, and when the transistor Qi is off 
the base of the transistors Q2, Q3 will have the voltage of the collector that is high. 
Having the voltage connected to the base of the transistors Q2 and through the 
resistor R'2 will turn the transistor Q2 on and turn the transistor off When the 


transistor Qi is turned on, the voltage will be connected to the gate of the synchronous 
MOSFET as in Figure IV.2(c). Now, the gate-source voltage of the synchronous 
MOSFET is equal to the capacitor voltage and the MOSFET will be turned on as in 
Figure IV.2(d). 

When < 0, either when the primary switch is closed or the current reversed 
its direction in case of discontinuous operation, the diode will be reversed biased as in 
Figure IV.2(e), and the base-emitter of the transistor Qi will be connected to through 
the resistor This will turn the transistor on, and the base of the transistors Qi and Qj, 
will be connected to a low voltage. This will turn the transistor off- ^^id the transistor 
Qj, will be turned on. The gate of the sjoichronous MOSFET will now be connected to a 
low voltage and it will be turned off as in Figure IV.2(f). 


Figure IV.2 Synchronous Driving Circuit Operation. 




Figure IV.3 Non-inverting Hysteresis Controller. 

Hysteresis Controller 

To be able to maintain the MFC voltage at the desirable level, the hyteresis 
controller can be used. The inverting hyteresis controller was used with the boost 
converter [5, 16]. When using the inverting hyteresis controller, a transistor must be 
added to invert the output signal. This transistor can be eleminated if the non-inverting 
hyteresis controller is used. The non- inverting hyteresis controller shown in Fgure IV.3 
can maintain the voltage of the MFC at the desirable level. Choosing the non-inverting 
hyteresis controller will reduce the number of elements used in the controller without 
affecting its function. 

The resistors R2 and R4. in Figure IV.3 are chosen to be variable resistors to 

change the hysteresis band. Using these variable resistors the voltage of the MFC and the 


operation frequency can be confroUed. The high threshold voltage and the low threshold 
voltage determine the hysteresis band. To calculate the values of the Kf/i-// ^th-u the 
following equations can be used: 

_ + R2) i?4 

^ (R, + R2) R4 _ KcRi 

(7?4+i?3)«2 R2 

when the MFC voltage hits the Kth-H, the controller turns the MOSFET on, and when the 
MFC voltage hits the Vtn-i, the confroUer turns the MOSFET off. 

System Simulation 

After choosing the self-synchronized flyback converter and the hj^eresis 
controller to harvest the energy fom the MFC, a simulation must be done to make sure 
that this combination will work with the MFC. The MFC was simulated as a battery with 
a series resistance as an internal resistance. The code was used to simulate the system is 
in the appendix and the results are shown in Figure rV.4. 

The first waveform in Figure IV.4 is the primary MOSFET gate signal, which is 
confroUed by the hj^eresis band. The second waveform is the voltage of the primary 
winding of the fransformer and notice that it is different than the MFC voltage because 
the primary inductor is disonnected from the MFC for part of the time depending on the 
duty ratio. The third waveform represents the primary and the secondary currents and 
notice that the secondary value depends on the turn ratio of the transformer. 



e 0.5 



: ^ON 






4 5 

Tim e [ma ec] 


Figure IV,4 Simulation Results. 

Overall System 

The overall system schematic of the proposed MFC energy harvester is shown in 
Figure IV. 5. To evaluate the system efficiency, a boost converter has been tested under 
the same conditions for comparison. 

At the beginning the hysteresis controller will turn the primary MOSFET on, and 
the current will flow through the primary winding. The voltage of the MFC will decrease 
as the current increase, and when the MFC voltage hits the lower threshold voltage the 
hysteresis controller will turn the primary MOSFET off. The synchronous gate drive will 
turn the synchronous MOSFET on and off depending on the current direction on the 
secondary side of the transformer. 


Figure IV,5 Overall System, 



Self-Synchronous Flyback Converter 

The main purpose of using the self-synchronous flyback converter is to increase 
the efficiency of harvesting the energy from the MFC. To get high efficiency, the 
parameters of the self-synchronous flyback converter must be chosen carefully, so they 
consume less energy. 

System Parameters 

To build the circuit, the parameters must be chosen such that they have low power 
losses. For example, the MOSFETs must have low on-resistance and the comparator in 
the hyteresis controller must consume minimal power. For harvesting the energy all the 
resistances in the path of the current must be reduced to reduce the amount of the power 

The flyback transformer was made using the transformer-winding machine shown 
in Figure V.l. The primary inductance is an important element because it affects the 
switching frequency. For this experiment it was chosen to have low inductance to reduce 
the number of turns, which will reduce the resistance of the wire. A 50-tum primary 
inductor was made using the transformer-winding machine, which has an inductance of 
7mH and 3.2D. resistance. Also, the turns ratio is important because as the turns ratio 
increase the secondary current will decrease. Reducing the secondary current should 
reduce the secondary losses. This can be seen from the simple Pio^s equation: 


Ploss — R 

If the secondary winding is doubled, the resistance will also be doubled. But the 
current will be reduced in half. Inserting those values in the loss equation will result in 
reducing the loss by half So, as the turns ratio increases the efficiency should increase. 
For this experiment the turns ratio was 1 :4. 

The MOSFETs should have low on resistance to reduce the power loss. The ON 
Semiconductor N-channel MOSFET 4906NG was used on both switches, which has 
6.5mn on-resistance at 4.SV Vcs- The PN2222A NPN & PN2907A PNP transistors were 
used for the transistors on the synchronous driving circuit. Also, the diode 1N755A was 
used as in the synchronous driving circuit. The resistor R[ must be very high to limit 
the current that flow through it, just enough to drive the transistor Qi- The resistor i?2 
should be high enough so that the current is limited when the transistor Qi is on. The 
values of the parameters that were used in this experiment are listed in table V.l. 


Figure V.l Winding Machine. 


Table V,l Table of Parameters, 



Traesformer Turns Ratio 

(1:4) (50 turn:400 turn) 

Transformer Priman^^ Induct.ance 

7 mH 






500 n 


300 Kn (Variable) 

10 kQ 

& feil (Variable) 


1 kn 

1 MO 

100 KQ 


Qi & Q2 



Output Capacitor 

1 F, 2.5 V 


DO 0.1 HA 0.6 


OS — ' — r ^ - 

0,2 • ■ 1 ■- 

OjQ D.2 da (L6 OM 

TIum (ms) 


Figure V,2 Ringing Waveforms, 

(a) Switch State (b) MFC Voltage. 

Filtering the MFC Voltage Ringing 

The circuit was buih in the lab, using the parameters chose on the previous 
section. When the built circuit was applied to the MFC and connected to the hysteresis 
controller, the voltage of the MFC started ringing as the switch goes off, and it becomes 
normal when the switch is on. This ringing is caused by the weakness of the MFC and it 
affected the hysteresis controller, since the voltage of the MFC exceeds the high and the 
low threshold voltages many times during the off period. This makes the hysteresis 
controller open and close the switch many times during the off period. This ringing on the 
MFC voltage and the hysteresis controller output waveforms can be seen in Figure V.2. 


Figure V.3 Non-inverting hysteresis controller with capacitor at the input. 

The MFC voltage ringing starts when the switch goes off, which means when the 
current stops flowing through the transformer primary winding. When the switch is off 
the current stops flowing and because the MFC is a weak source the voltage starts 
ringing. This problem increases since the hysteresis controller opens and closes the 
current path many times, which will make the ringing worse. 

To solve this problem, a capacitor must be connected at the input of the hysteresis 
controller as shown in Figure V.3. Now when the switch is off 
and the path of the current is closed, the current will flow through the capacitor and 
charge it. When the switch is on, the capacitor will be discharged and the current will be 
added to the current from the MFC. The waveforms of the MFC voltage and the 


4JJ T 

lime \ms) 

a* T 



f 0,3 

□ 25 --■ 






IliM (ma) 


Figure V,4 Waveforms After Filtering. 

(a) Switch State (b) MFC Voltage. 

MOSFET gate signal from the hysteresis controller after applying 0.1/uF capacitor are 
shown in Figure V.4. 


The proposed system was built as shown in Figure V.5 and applied to the MFC 
for 25 minutes and the energy was harvested and stored in the output capacitor. The 
readings were taken each 5 minutes for the MFC voltage, MFC current, output capacitor 
voltage, switching frequency and exported to excel for the calculations and they can be 
seen in Table V.2. The resultant waveforms were recorded using Tektronix TPS2012 
oscilloscope as shown in Figure V.6. 




After getting the experiment results, the efficiency must be calculated. To 
calculate the efficiency the input and the output energy must be calculated because we 
know that the efficiency is equal to: 

where the input energy can be calculated using: 

Where and Ii-n are the input voltage and input current respectively. Now we have the 
input energy, which is the energy extracted fi-om the MFC. The output energy in our 
experiment is stored in the output capacitor, and the energy stored in the capacitor can be 
calculated fi-om: 

Where is the capacitor voltage and C is the capacitance of the capacitor. 

The efficiency was calculated for every 5 minutes for this experiment and was 
plotted in Figure V.7. The average efficiency was 46.1%, and from Figure V.7, it is clear 
that the efficiency at the beginning is low because at the beginning the diode conducts, 
and when the output capacitor voltage reached 1.27 V the synchronous driving circuit 
started to work and the efficiency started increasing. As the output capacitor voltage 
increases the gate voltage will increase, which will turn the sjTichronous MOSFET on. 
The on-resistance of the MOSFET depends on the voltage at the gate. The on-resistance 


Table V.2 Self-Synchronous FLyback Converter Experiment Results, 

Time [min] 


Frequency [KHz] 


























decreases as the gate voltage goes high, and the MOSFET is completely turned on when 
the gate voltage is equal to the rated gate-source voltage on the data sheet. 

The switching frequency is decreasing because of the capacitor across the input of 
the hysteresis controller. When the primary switch is off the current from MFC charges 
the filtering capacitor, which makes the MFC voltage need more time to hit the upper 
threshold of the hysteresis controller. 


1 T 

1 - 



1.C iS 
Tlnw (imj 






1 f 




—I — 

— 1— L 


— 1 

1.0 IS 

Tlmi (fTu| 


Figure V.6 Self-Synchronous FLyback Converter Experiment Waveforms, 

(a) Primary Switch State (b) MFC Voltage (c) Synchronous Switch State. 


SO.ODft ■ 

■i 30.00% ■ 



lO.OOS ' 



Time (min) 



Figure V.7 Efficiency of the Self-Synchronous Flyback Converter vs. Time. 


Table V,3 Boost Converter Experiment Results, 

Time [mini 

L J 

Ijnfr [inAl 

C L J 

Frequency [KHzl 


























Boost Converter 

To compare the results of harvesting the energy from the MFC with the self- 
synchronous flyback converter, diode based-boost converter will be used. Comparing the 
results of the two converters will help to see the improvement of the self-synchronous 
flyback converter in terms of the efficiency. To make the comparison fair, the two 
converters have similar components. The inductor used with the boost converter was 
made in the lab using the winding machine. This inductor has 7mH inductance and 3.2fl 
resistance. Although this inductor is not efficient because of the high resistance compared 
to the inductor used in [5], it was used because we will use the same winding machine to 
make the transformer of the fiyback converter. Low on-resistance 4906NG MOSFET has 
been chosen with the 1N755A diode to build the boost converter. The three main 
components for the boost converter were chosen and the circuit was built and tested in 
the lab. 


Figure V,8 Boost Converter Experiment Waveforms. 

(a) Switch State (b) MFC Voltage 

The boost converter was connected to the MFC with the non-inverting hysteresis 
controller, and the energy was harvested and stored in 1 F supercapacitor. The 
experiment took 25 minutes and the readings were taken every 5 minutes. The MFC 
voltage, the MFC current, the capacitor voltage, and the switching frequency were 
recorded and they can be seen in table V.3. The waveforms shown in Figure V.8 were 
taken using Tektronix TPS2012 oscilloscope. The voltage of the MFC can be seen 
controlled by the hysteresis band of the hysteresis controller. The results was exported to 
excel and using the same calculations for the self-synchronous flyback converter, the 
efficiency was calculated and plotted as shown in Figure V.9, and the overall efficiency 
was calculated to be 33.5 %. 


As the output capacitor voltage increase, the switching frequency increase 
because of the higher resistance for injecting the power and this increase in the switching 
frequency can be seen in Table V.3. 



Time (tnin) 



Figure V.9 Efficiency of the Boost Converter vs. Time, 




The two converters were operated at the same conditions to make the 
comparison fair. The MFC voltage is shown in Figure VI. 1 in real time for both 
converters. The MFC current is also shown in Figure VI. 2 for both converters in real 
time. Having the same voltage and current from the MFC means the same input power 
for both converters. The thing that is different between the two circuits is the switching 
frequency. They started at approximately the same switching frequency, but the 
switching frequency of the flyback converter was decreasing and was increasing for the 
boost converter. This switching frequency behavior difference is shown in Figure VI.3. 

The self-sjoichronous flyback converter charged the ou^ut capacitor faster than 
the boost converter and Figure VI.4 shows a comparison between them. It is clear from 
the efficiency comparison in Figure VI. 5 that the self-synchronous flyback converter has 
higher efficiency than the boost converter after 10 minutes of operation, which is because 
the synchronous driving circuit started to drive the synchronous MOSFET at this time. 
Even the synchronous driving circuit started to drive the synchronous MOSFET after 
approximately 10 minutes, the overall efficiency was improved by 37.6% to reach 46.1% 
compared to 33.5% for the boost converter. The self-synchronous flyback converter was 
able to store 2.27J out of 4.91J in the output capacitor compared to 1.665J out of 4.95J 
stored in the output capacitor by the boost converter. 










L~i — 

— i — 

• BODiSt 

W flyback 


IC II 20 

Time (mm]- 


Figure VI, 1 MFC Voltage for Both Experiments vs. Time. 



3 ^ 




Time (mm) 



Figure VI,2 MFC Current For Both Experiments vs. Time, 


Figure VI.3 Switching Frequency For Both Experiments vs. Time, 


O.MK 1 1 1 1 

S 10 IS 20 25 

Time [mm] 

Figure VI,5 Efficiency of Both Converters vs. Time. 



The self-sjoichronous flyback converter was designed and built, and the energy 
was harvested from the MFC and stored in the capacitor in a good efficiency compared to 
the boost converter. The efficiency improved when the secondary diode was replaced by 
a MOSFET, because the diode has a high voltage drop across it (0.7 V). Replacing the 
diode by the MOSFET resulted in floating switch, which needs to be driven by an 
isolated source. The synchronous driving circuit [20] was used to drive the sjoichronous 

The main advantage of using the DC-DC converters is the ability to control the 
voltage of the MFC. The non-inverting hysteresis controller was used for this thesis, but a 
capacitor was needed in the input of the hysteresis controller to filter the voltage of the 
MFC fi-om the ringing caused by closing the current path each cycle, which is the way 
that the flyback converter works. 

This thesis proved that replacing the secondary diode by a MOSFET improved the 
eciency of harvesting the energy from the MFC. It also proves the advantages of using 
DC-DC converters, which are the ability to control the MFC voltage, the ability to store 
the energy in the output capacitor, and the efficient way to harvest the energy from the 



This is the simulation code for the flyback converter: 

clear all; 
close all; 

screenSize = get(0, ' ScreenSize ' ) ; 

position = [ screenSize { 3 ) /3 screenSize ( 4 ) /5 560 420]; 
set ( , ' Def aultFigurePosition ' , position ) ; 

tMax =0.01; 
C = 40000e-6; 
dt = le-6; 

% time period [sec] 

% Sampling time [sec] 

Vint = 0.7; 
R = 120; 
R2= 0; 

% Input voltage pealc [V] 
% Resistance [Ohm] 

t = linspace ( , tMax, tMax/dt ) 

vs = Vint*ones(l, length(t)) 

vL = zeros(l, length(t)); 


vR = zeros(l, length(t)); 


vR2 = zeros(l, length(t)); 

vo = 0.6*ones(l, length(t)); 

11 = zeros(l, length(t)); 

12 = zeros(l, length(t)); 
sw = ones(l, length(t)); 

% time vector 

% Source voltage vector 

% Inductor voltage vector 

% Resistor voltage vector 

% Current vector initialization 

% load current 

N = 4; 

Ll= 200e-3 ; 
L2= LI * N ; 
Th_H = 3 /lOOO ; 
Th_L = 2/N/lOOO ; 

for n = l:length(t)-l 
if sw(n) == 1, 

vR(n+l ) = R * il (n) ; 

vL(n+l) = vs(n)-vR(n); % Inductance voltage 

il(n+l) = il(n) + (vL(n) ) /Ll*dt; 

vo(n+l) = vo(n) + (- il)/C*dt; 

if il(n) >= Th_H, 
12 (n+1 )=il (n+1 ) /N; 





vR2 (n+1 )= R2 * 12 (n) ; 
vL(n+l)=(vR2(n+l)-vo(n+l) )/N; 

12 (n+1) = 12 (n) -(vo(n)/L2) * dt; % Integration for current 

vo(n+l) = vo(n) + ( 12 ( n ) -11 ) /C*dt ; 

vL(n+l)= (-vo(n+l) )/N; 

If 12 (n) <= Th_L, 
sw(n+l) = 1; 
11 (n+1 )=12 (n) *N; 




tm = t*1000; 
% Result plotting 
figure ( 1 ) ; 
subplot (3,1,1); 
plot ( tm, sw, ' k ' ) ; 
grid on; 

axls([0 tMax*1000 1.5]); 
ylabel( ' Switch Status ' ) ; 

subplot ( 3 ,1,2); 
plot ( tm, vL, ' k ' ) ; 
grid on; 
hold on; 

ylabel (' Primary Inductor Voltage [V] ' ) ; 

subplot (3,1,3); 
plot ( tm, 11 , ' k ' ) ; 
grid on; 
hold on; 

plot ( tm, 12 , ' r ' ) ; 
ylabel (' Current [mA]'); 
xlabel('Tlme [msec]'); 

figure ( 2 ) ; 
plot(tm, vo) ; 
hold on; 
grid on; 


sw( n+1 ) 
11 (n+1 ) 

= 0; 


1. Conrad Donovan, Alim Dewan, Deukhyoun Heo, and Haluk Beyenal,Batteryless, 
Wireless Sensor Powered by a Sediment Microbial Fuel Cell, Environmental 
Science & Technology 2008 42 (22), 8591-8596. 

2. Shantaram, H. Beyenal, R. Raajan, A. Veluchamy and Z. Lewandowski,Wireless 
Sensors Powered by Microbial Fuel Cells, Envi- ronmental Science & 
Technology 2005 39 (13), 5037-5042. 

3. S.-E. Oh, B.E. Logan, Voltage reversal during microbial fuel cell stack operation, 
Journal of Power Sources, Volume 167, Issue 1, 1 May 2007, Pages 11-17. 

4. P. Aelterman,K. Rabaey,H. The Pham,N. Boon, and W. Verstraete,Continuous 
Electricity Generation at High Voltages and Currents Using Stacked Microbial 
Fuel Cells, Environmental Science & Technology 2006 40 (10), 3388-3394. 

5. Jae-Do Park; Zhiyong Ren ,"E_cient energy harvester for microbial fuel cells 
using DC/DC converters," Energy Conversion Congress and Exposition (ECCE), 
2011 IEEE , vol., no., pp.3852-3858, 17-22 Sept. 2011. 

6. Jae-Do Park, Zhiyong Ren,Hysteresis controller based maximum power point 
tracking energy harvesting system for microbial fuel cells. Journal of Power 
Sources, Volume 205, 1 May 2012, Pages 151-156. 

7. Woodward, L., Perrier, M., Srinivasan, B., Pinto, R. P. and Tartakovsky, B., 
(2010), Comparison of real-time methods for maxi- mizing power ou^ut in 
microbial fuel cells. AIChE J., 56: 27422750. 

8. Komeel Rabaey, Willy Verstraete,Microbial fuel cells: novel biotechnology for 
energy generation. Trends in Biotechnology - 1 June 2005 (Vol. 23, Issue 6, pp. 

9. R.A. BuUen, T.C. Arnot, J.B. Lakeman, F.C. Walsh, Biofuel cells and their 
development. Biosensors and Bioelectronics, Volume 21, Issue 11, 15 May 2006, 
Pages 2015-2045.47 

10. Alim Dewan, Haluk Beyenal, Zbigniew Lewandowski, Scaling up microbial fuel 
cells. Environmental Science & Technology 2008 42 (20), 7643-7648 

11. Meehan, A.; Hongwei Gao; Lewandowski, Z., "Energy Harvesting With 
Microbial Fuel Cell and Power Management System," Power Electronics, IEEE 
Transactions on, vol.26, no.l, pp. 176,1 81, Jan. 2011. 


12. R.P. Pinto, B. Srinivasan, S.R. Guiot, B. Tartakovsky,The effect of real-time 
external resistance optimization on microbial fliel cell performance, Water 
Research, Volume 45, Issue 4, February 201 1, Pages 1571-1578. 

13. Alim Dewan, Conrad Donovan, Deukhyoun Heo, Haluk Beyenal,em Evaluating 
the performance of microbial fuel cells powering electronic devices. Journal of 
Power Sources, Volume 195, Issue 1, 1 January 2010, Pages 90-96. 

14. Fei Zhang, Lei Tian, Zhen He,Powering a wireless temperature sensor using 
sediment microbial fuel cells with vertical arrangement of electrodes. Journal of 
Power Sources, Volume 196, Issue 22, 15 November 201 1, Pages 9568-9573. 

15. Heming Wang, Zhiyong Ren, Jae-Do Park,Power electronic converters for 
microbial fuel cell energy extraction: Effects of inductance, duty ratio, and 
switching frequency. Journal of Power Sources, Volume 220, 15 December 2012, 
Pages 89-94. 

16. Jae-Do Park, Zhiyong Ren,High efficiency energy harvesting from microbial fuel 
cells using a synchronous boost converter. Journal of Power Sources, Volume 
208, 15 June 2012, Pages 322-327. 

17. Richelli, A.; Colalongo, L.; Tonoli, S.; Kovacs, Z.,"A 0.2V1.2V converter for 
power harvesting applications," Sohd-State Circuits Conference, 2008. ESSCIRC 
2008. 34th European , vol., no., pp.406, 409, 15-19 Sept. 2008. 

18. Lhermet, H.; Condemine, C; Plissonnier, M.; Salot, R.; Audebert, P.; Rosset, 
M., "Efficient Power Management Circuit: From Thermal Energy Harvesting to 
Above-IC Microbattery Energy Storage," Solid-State Circuits, IEEE Journal of , 
vol.43, no.l, pp.246,255, Jan. 2008. 

19. Carlson, E.J.; Strunz, K.; Otis, B.P.,"A 20 mV Input Boost Converter With 
Efficient Digital Control for Thermoelectric Energy Harvesting," Solid-State 
Circuits, IEEE Journal of, vol.45, no. 4, pp. 74 1,750, April 


20. Lee Jong-Jae; Jung-Min Kwon; Kim Eung-Ho; Woo- Young Choi; Bong-Hwan 
Kwon," Single- Stage Single-Switch PFC Flyback Converter Using a Synchronous 
Rectifier," Industrial Electronics, IEEE Transactions on , vol.55, no.3, 
pp.1352,1365, March 2008. 

21. Liu, H. and Logan, B.E. (2004) Electricity generation using an aircathode single 
chamber microbial fuel cell in the presence and absence of a proton exchange 
membrane. Environ. Sci. Technol. 38, 4040-4046. 


22. Schro"der, U. et al. (2003) A generation of microbial fuel cells with current 
outputs boosted by more than one order of magnitude. Angew. Chem. Int. Ed. 
Engl. 42, 2880-2883. 

23. Rhoads, A., H. Beyenal, and Z. Lewandowski. 2005. Microbial fuel cell using 
anaerobic respiration as an anodic reaction and biomineralized manganese 
as cathodic reactant. Environ. Sci. Technol.39: 4666-4671. 

24. Liu H, Cheng S, Logan BE. Production of electricity from acetate or butyrate 
in a single chamber microbial fuel cell. Environ Sci Technol. 2005;39:658-